Can physical exercise modulate cortisol level in subjects with depression ? A systematic review and meta-analysis

INTRODUCTION
Major depressive disorder (MDD) is a prevalent medical illness that is associated with chronic stress. Patients with MDD can show an imbalance in cortisol levels, which can be restored with the remission of symptoms. Physical exercise training has been used as a tool to promote changes in cortisol levels in healthy individuals. However, it is unknown if exercise can produce the same results in individuals with MDD.


OBJECTIVE
To review evidence of cortisol changes after exercise training in individuals with MDD.


METHODS
We conducted a search on PubMed, SciELO, LILACS, ISI Web of Knowledge, Scopus, and PsycInfo databases. Random effects meta-analysis was performed and standardized mean difference (SMD) effect size was calculated. Analyses of forest and funnel plots was conducted using Stata v.11.0 software.


RESULTS
At first, 463 studies were obtained in the search. After completion of the selection procedure, five articles with seven analyses were included. Type of exercise, frequency of training, cortisol measurement, and type of control group were analyzed. There was a reduction of cortisol levels in the exercise group (SMD = -0.65, 95%CI 1.30-0.01). Moreover, sensitivity analysis and subgroup analyses revealed an effect of type (aerobic exercise) and frequency (five times per week) of exercise on reduction of cortisol levels. However, these results should be interpreted cautiously due to the small number of studies and a substantial heterogeneity among them.


CONCLUSION
Physical exercise promotes a reduction in cortisol levels in individuals with MDD. However, this finding can be influenced by type of exercise, weekly frequency, and type of cortisol measurement.


Introduction
Major depressive disorder (MDD) is a prevalent medical illness that comprises a variety of different symptoms (e.g., depressed mood, anhedonia, changes in appetite and sleeping, loss of energy, somatic symptoms and others). 1 The pathophysiology of MDD is related to several neurobiological changes, including hypothalamic-pituitary-adrenal (HPA) axis malfunction and consequent increase [2][3][4] or decrease [5][6][7] in basal cortisol levels.[16] A meta-analysis reinforced that physical exercise is an effective non-pharmacological treatment for depression in older adults. 17There are several physiological hypotheses to explain the positive effect of exercise on MDD, including HPA changes. 18,19Physical exercise can modulate cortisol levels after a single session of exercise or after physical exercise training, 20 possibly due to an upregulation of glucocorticoid receptor. 21However, it is still unknown whether exercise training might promote the same cortisol effects in individuals with MDD.
A previous article suggested that mindful exercise involving body movement and breathing can reduce stressful signals to the limbic system and thus HPA activity, resulting in reduced release of adrenal glucocorticoid. 22Likewise, this hypothesis might be extrapolated to other types of exercise.A previous study from our laboratory showed that the differences in basal cortisol between MDD and healthy control groups can be modulated by physical performance aspects. 23However, although the association between HPA malfunction and MDD has been investigated for five decades, 24 few interventional studies have evaluated the effect of physical exercise on cortisol levels in individuals with MDD, and the few results available are contradictory.[33] Therefore, the aim of the present study was to systematically review the literature investigating the effects of physical exercise training on cortisol levels in subjects with MDD and to explore potential moderators of this relationship.It is expected that physical exercise has an important effect on cortisol levels in individuals with MDD.

Protocol and registration
This study was registered on the International Prospective Register of Systematic Reviews (PROSPERO) under protocol no.CRD42018055996.

Eligibility criteria
The present systematic review and meta-analysis was conducted according to the PICO search strategy 34 and the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement. 35

Search
The search strategy was performed using a combination of the keywords "exercise" AND "cortisol" AND "depression" with the descriptors "physical activity" AND "major depression". of the studies selected were evaluated independently and the meta-analysis was performed.

Data collection process
Two authors (AHNB, HSM) extracted the data from each study and entered them into a table with the following information: sample (age group, sex, clinical diagnosis), exercise method (type, intensity and duration of physical activity), cortisol method (type and time of collection), cortisol levels, and depressive symptoms.

Risk of bias in individual studies
Risk of bias was assessed according to the Physiotherapy Evidence Database (PEDro) scale, 36 considering the following items: eligibility criteria; subjects were randomly allocated to groups; allocation was concealed; the groups were similar at baseline regarding the most important prognostic indicators; there was blinding of all subjects; there was blinding of all therapists who administered the therapy; there was blinding of all assessors who measured at least one key outcome; measures of at least one key outcome were obtained from more than 85% of the subjects initially allocated to groups; all subjects for whom outcome measures were available received the treatment or control condition as allocated or, where this was not the case, data for at least one key outcome was analyzed by intention to treat; the results of between-group statistical comparisons are reported for at least one key outcome; and the study provides both point measures and measures of variability for at least one key outcome The score of this scale can range from 0 to 10 (yes = 1; no = 0).The first item was not included in the total score.Studies were considered to be of good quality when a score of 6 to 8 was reached, of moderate quality when the score was 4 to 5, and of poor quality when the score was below 3.

Statistical analysis
We performed a random effects meta-analysis and estimated the effect size using standardized mean differences (SMD).

Study selection
At first, 463 studies were obtained in the search considering exercise, cortisol, and depression.Following application of the eligibility criteria, 24 studies were selected for full-text reading and review.After the selection procedure, five studies, with seven arms, were included in the meta-analysis.The selection process of the five studies 25,26,29,30,39 included in this meta-analysis is showed in Figure 1.

Study characteristics
Overall, the five studies included 209 participants All the studies included were randomized controlled trials (Table 1).The duration of the intervention varied from 8 to 16 weeks for chronic studies.Different types of exercise were used, including aerobic exercise, 25,29,30 Records jogging training, 26 strength training, 29 relaxation exercise, 29 Buddhism walking meditation, 37 traditional walking 39 and stretching. 25ercise intensity was also different among the studies.One study assessing strength training investigated 75% of repetition maximum (RM), 29 while the studies employing aerobic training used exercises of both high intensity 29 (n = 1) and mild to moderate intensity (n = 4). 25,26,30,39rtisol measurement was made in the saliva (n = 1), 25 urine (n = 1), 26 and plasma (n = 3) 29,30,39 of the participants, at different times of the day.Cortisol assessments were also different among the studies.
Two of them evaluated cortisol in basal condition, 29,30 one article observed cortisol awakening response, 25 another study assessed cortisol levels after 8 hours of overnight fasting 39 and the last one measured 24 hours of excretions. 26

Risk of bias across studies
Visual analysis of the funnel plot showed asymmetry of two articles 26,29 (three trials) within the studies,

Discussion
The current systematic review and meta-analysis aimed to synthesize the data available on cortisol changes after exercise training in individuals with MDD compared to control groups.A moderate reduction of cortisol levels was found following an exercise intervention in people with depression.However, these results should be interpreted cautiously due to the small number of studies and a substantial heterogeneity among them considering type of exercise, type of cortisol measurement, frequency of exercise, and type of control group.
Related to the type of exercise, the great majority of studies investigated aerobic exercise. 25,26,29,30,39rength training 29 and holistic exercise (walking meditation) 39 were also evaluated.Although a higher effect was observed for aerobic training, this result did not reach statistical significance, probably due to the small number of articles included in this meta-analysis.
Moreover, the results were divergent.One study 29 observed non-significant reduction of cortisol after both aerobic and strength training.Another study 30 also observed non-significant differences in cortisol levels.
It is interesting to note that these studies investigated cortisol at baseline conditions.However, studies which have observed cortisol awakening response, 25 morning cortisol after overnight fasting, 39 and 24 hours of excretion through urine 26 have found significant cortisol reductions.Since changes in diurnal rhythms of cortisol are well known to occur in some cases of MDD 40 and can vary according to time of sampling, 3 cortisol should be measured at different points of the day or in response to stimuli (awakening response or stressor) to produce remarkable results.
Another possible reason to explain the wide variability of the results is the type of specimen collection used to measure cortisol.Although a previous meta-analysis has concluded that this factor does not appear to impact the magnitude of change in cortisol levels prevs.post-treatment, 41 we observed a very large effect for urine, following by a moderate effect for saliva and a small effect for plasma.Plasma measurement is a widely used method, however it is the most invasive form when compared to saliva and urine collections (i.e., a stressful procedure).It is possible that 24hour urine measures have the advantage of providing integrative HPA axis measures over larger time periods.
Nonetheless, compliance is poor for 24-hour measures, and renal conditions may affect urinary cortisol levels. 42n our meta-analysis, only one study evaluated this method.
Especially with regard to exercise methodology, the studies were also very discrepant, which may have impacted the results.For example, three different types of weekly frequencies were used: two, 29 three 25 and five times per week. 26In our study, a very large effect was found for training when the frequency was five times a week, whereas a moderate effect was present for three times a week and a small effect for two times a week.Other parameters should be considered, such as exercise intensity, which can impact the weekly calorie expenditure and promote greater effects in depressive symptoms. 43However, due to the heterogeneity of intensity measurement, we were not able to specifically evaluate dose response.
The variability of activities in the control groups was also an important methodological problem observed in the studies.For example, two studies 25,29 investigated an active control group (stretching and relaxation, respectively), while three studies 26,30,39  Some limitations and strengths of the current metaanalysis should be considered.First, it is not possible to conclude that the effect of exercise on cortisol in individuals with MDD occurs concomitantly with reductions in depressive symptoms.Among the studies here evaluated, three showed significant reductions of depressive symptoms.However, one study did not present the association between reduced cortisol and improved depressive symptoms, 30 and another found non-significant results. 29Second, our meta-analysis involved patients with a clinical diagnosis of depression and individuals with depressive symptoms according to depression scales.Due to the small number of articles included and the methodological differences among them, we cannot conclude if there was an influence of clinical diagnosis in the effect of exercise on cortisol levels.Finally, the small number of articles did not allow to reach a reliable conclusion regarding the subgroup analysis performed.Fortunately, most of the studies assessed in this meta-analysis were classified as good quality according to risk of bias within the studies.
However, we assume that the quality of the trials published should be analyzed more cautiously.
Even though the findings of the present meta-analysis were not in accordance with the initial hypothesis, it is important to highlight that the modulation of cortisol in response to exercise training for depression needs to be investigated in more detail.
In conclusion, physical exercise promotes a moderate reduction in cortisol levels in MDD individuals.However, this finding can be influenced by methodological aspects, such as type of exercise, weekly frequency, and type of cortisol measurement.

(
125 intervention vs. 84 control subjects).Three of the five studies used clinical diagnosis based on DSM-IV or ICD-10 criteria, and the other two investigated depression using scales (GDS and CES-D).The mean age of the individual populations varied from 18.7±0.2to 81.0±1.7 years old.

Table 1 -
Summary of studies